论文部分内容阅读
确定多元医学参考值的单侧的统计方法中,有多指标正态分布法、多指标百分位数法、多维标度法与组合指标法。这些方法要求各指标医学意义的效应相同或等价诊断意义。对于观测指际具有不同的诊断意义,则上述方法不能使用。对于两类指标,不仅某一类单独出现异常时有意义,且各类均有异常时亦有意义,并考虑两类指标有不同阳性率情况,根据某受试者的异常指标愈多,愈有可能患相应的某病,以医学意义限定在不同情况下最少异常指标数,按概率论提出两类指标组合指标法。文中有医学应用实例、公式证明与讨论。
Among the unilateral statistical methods used to determine the multiple medical reference values, there are multi-indicator normal distribution method, multi-indicator percentile method, multi-dimensional scale method, and combination index method. These methods require the same medical significance of the indicators or equivalent diagnostic significance. The above methods cannot be used because they have different diagnostic significance for observational indicators. For the two types of indicators, not only is it significant when a particular type of abnormality occurs, but also when there are abnormalities of each type, and considering the different positive rates of the two types of indicators, the more abnormal indicators according to a certain subject, the more There may be a corresponding disease, with the medical sense limiting the number of abnormal indicators in different circumstances, according to the probability theory put forward two types of indicators combined index method. There are medical application examples, formulae proof and discussion in the text.